Pharma development and manufacturing with QbD 2.0

Posted: 3 July 2014 | | No comments yet

Pharma and BioPharma industries are aware of the impact of production processes on sustainability of business operations. To improve performance, companies have recognised that it is necessary to better understand the drivers of both costs and revenues and the actions that can be put in place to address them.

In the past, commercial manufacturing emphasis was on full compliance with initially established product specifications leading to a perception of quality assurance based on testing, and avoiding later changes after regulatory submission. Although final product testing is an important element of quality control, final product quality can be measured but not modified, leading to product rejection or reprocessing activities and underperforming business outcomes, two kinds of waste according to lean manufacturing.

A body of documents from regulatory authorities framing the use of Quality-by-Design (QbD) elements exists. Namely, ICH guidelines (Q8-Q11) and FDA guidance (viz., the lifecycle process validation from 2011) all promote a science-based and data-driven approach – based on prior knowledge and new process understanding obtained through PAT tools (Process Analytical Technologies) – both in the development and manufacture of new or existing drug substances and products of either chemical or biological origin.

For QbD to materialise, particularly in regulated industries such as Pharma and BioPharma, PAT tools need to be used as in-process tools, preferably on whole samples, enabling their use as process state estimation techniques and not as one more way of monitoring parameters in a sample. Exploiting the intrinsic process fingerprinting capabilities of PAT tools will enable any disturbances on process conditions or on the feed stocks used as raw materials, to be detectable and tightly controlled while processing is still ongoing.

Pharmaceutical processes are sequences of unit operations. As such, raw-material attributes can be linked to processing conditions and end-product specifications. In current QbD very little is done by such interactions or even through a holistic end-to-end perspective to manage all critical sources of variability in a feedforward way (Figure 1).

However, another level of integration exists: across process/product lifecycle. From process/product design, development, industrialization and commercial manufacturing there should be both a rationale and a traceable sequence of events and elements, which can for instance be used to troubleshoot or improve that same process/product in the future.

An updated version of the QbD initiative (i.e., QbD 2.0) is therefore necessary to demonstrate how it can be applied not only to the manufacturing process, but as a holistic approach from R&D up to production, including technology transfer and knowledge management activities across the lifecycle of different products and manufacturing processes within a same technology platform.

That requires a much larger set of tools and skills than those in current QbD, which can be grouped under the name Manufacturing Sciences and Technologies (MSAT), and which should be aimed to support technical operations and achieve manufacturing excellence over lifecycle and across similar technology platforms.

MSAT approaches are PAT and QbD-based and include in addition, process systems engineering (PSE) principles, Lean Six Sigma (L6S) for routine manufacturing, but also Design for Six Sigma (DFSS) to support process transfers across different manufacturing sites, manage knowledge within the organisation, optimise processes and their operations, establish process and data management automation, troubleshoot activities and implement new technologies to support operations, while maintaining the connection to business outcomes.

Gap Analysis

A generally accepted QbD workflow (Figure 2), based on an earlier sketch of Prof. Staffan Folestad (Astrazeneca, 2003 – personal communication) is described by ISPE (Pharm. Eng 2010, PQLI Guideline 2011), which now incorporates the lifecycle continued process verification aspects of modern validation missed earlier. However in agreement with that construct, it seems logical to consider different levels of knowledge integration, both horizontally (whole process analysis) and vertically (over process history or lifecycle, from development to manufacturing). These additional levels will require knowledge management capabilities to be defined within a specific product/process technology platform.

But where are we after 10 years of QbD/PAT implementation? We have not reached the mature stage where we fully understand and are able to manipulate processes and specific unit operations, taking into consideration their interactions to achieve consistent end-product quality. Pharma and biopharma manufacturers still lack the regulatory flexibility promised by the introduction of the design-space concept. Moreover the current design-space definition will be sub-optimal when interactions between processing steps are not taken into account (as of today). Processes are still largely run not taking into account upstream variability. Furthermore, although great investment has been made in quality control of raw material using PAT tools or laboratory analytics, the industry is not at a stage where it truly knows all sources of variability and how to keep them under control. Therefore at a plant level performance is non-optimal and the dream of QbD and Class-A operation has been eluded simply because QbD is currently formulated to address one-unit-operation at a time.

The use of PAT in Pharma and Biopharma industries in the last 10 years has been focused on monitoring intensification (Process Analytical Chemistry oriented) across the flow sheet. The industry has not been able to take full advantage of PAT and QbD implementation, as the focus on PAT should be centered in process state estimation and not as a sophisticated alternative to other analytical devices.

Another aspect disregarded is the integration of engineering principles. In fact, this might be one of the largest gaps within the pharmaceutical industry that needs to be quickly surpassed so that QbD can be established with its full potential. Not only are observability and state estimation overlooked at the unit operation level, also almost no use is made of the strong interconnected nature of pharmaceutical processing. As most operations in Pharma and Biopharma are batch and follow first order kinetics, they are dependent on initial conditions (i.e., raw material or intermediates attributes), and have end-points which, in many cases, can be predicted under GMP conditions from initial conditions. Thus, ‘first samples’ are determinant for state estimation as they set the unit operation trajectory, particularly when following a ‘recipe’.

In short, whole process samples more faithfully capture the process state (e.g. matrix with products and reactants contains information on stoichiometries) than QC of specific parameters after undergoing several analytical procedures. A good example is homogeneity in blends. A single sample is not representative of the process state but inline spectra can be used to determine content uniformity.

MSAT Strategies: plant wide excellence over the lifecycle

An end-to-end analysis connecting all process steps to end-product quality must be adopted comprehensively, in parallel to a time-wise integration of data, information and knowledge. It is reasonable to ask how PAT solutions are being managed over the product lifecycle and how companies are generalizing the strategies and knowledge accumulated within their product and technology portfolios?

For establishing such lessons learned from one case to the other Knowledge Management (KM) tools are required, to ensure consistency and clarity of QbD implementations. That requires standardization of procedures and activities, to ensure the capturing, storing and in future re-using of information and knowledge already available and stored across an organization. This would allow not only consistency of QbD filings and implementations across different products, but also technology platforms. So far, we are not yet there! The industry and its regulators are still locked in the current QbD formulation, oriented to bring QbD into one unit operation at a time, with the belief that improving each unit operation will improve global performance. This is not necessary true, since the best performance of each unit operation does not guarantee whole process reliability and robustness, neither product quality or consistency can be ensured at a six-sigma level. Practice shows that today unit operations are run and largely neglect the impact they have on further downstream processing.

This can be somehow avoided by incorporating L6S and DFSS methodologies from product design until routine manufacturing with an end-to-end perspective for achieving the QTPP very well set up by QbD 1.0, but delivered only by QbD 2.0. Although it is important to achieve Operational and Business Excellence, to our best knowledge no significant use of DFSS has been taken into account in the development of new products within QbD and PAT.

Unlike L6S, DFSS is not a strategy to improve the current processes with no fundamental change in process structure (“not firefighting”). It is a design or redesign methodology that is disruptive in its nature and as such can be considered for developing new products and processes and can be started on the earliest stages of the process life cycle or when a redesign of the product or process is required. It is focused on achieving Six Sigma from the design phase and as such prevents significant efforts of troubleshooting along the life cycle (‘fire preventing’).

Finally, it has to be taken into account that process development should consider that optimization efforts have to be taken on a level higher than the industry current practices, i.e., optimization of unit operations has to be considered simultaneously and not one at a time. As this is not possible within the current toolset, the Pharma and Biopharm industry should consider Plant Wide Optimization strategies, as discussed during the 80s in Class-A Chemical Engineering industry.

As Pharma and Biopharma greatly evolve their product and process development and life cycle management towards a systems engineering view they will need to incorporate tools to promote continuous process improvement with QbD and PAT aligned with Lean and Six Sigma activities and disruptive improvements aligning Design for Six Sigma, Operational Excellence and Knowledge Management and end-to-end systems thinking and overall holistic performance. The implementation of these MSAT strategies will allow companies to achieve better results earlier and throughout the life cycle of their products and processes (Figure 3).


We proposed some directions through MSAT approaches to achieve better global performance for modern pharma manufacturing using QbD. That approach must have an end-to-end over lifecycle aspect and will use a generic set of tools that can fall under Quality as manufacturing-sciences and several data-driven technologies. Using systems thinking of product and processes capable of integrating QbD and PAT within L6S activities for improvements of manufacturing operations and DFSS for future products and processes design or re-design is a must. The level of such integration will be one of major challenges the industry and its regulators will have to overcome in the years ahead. Such efforts will require tools that are not yet available or completely developed, such as knowledge-management platforms.


Jose C. Menezes has been a Bioengineering Professor at TU Lisbon since 2005, and an Invited Professor at the Faculty of Pharmacy Lisbon from 2009. He is Founder of The Lisbon Pharmaceutical Engineering MSc Program (2006 -), and focuses on Pharma & Biopharma Project Management in PAT & QbD (1996 -). He has written over 20 papers and given conference talks, graduated 12 PhDs and 60+ MScs theses. He is an Active member of ISPE, APV, PDA, ACS, AICHE and IFPAC.

Pedro G Felizardo, PhD in Chemical Engineering at TU Lisbon. He is CTO and Partner at 4Tune Engineering Ltd, and an expert in Lean-Six Sigma and GMP. He is responsible for several QbD implementations for large Pharma and Biopharma, with over 10 years of experience in Industrial PAT applications (from raw-material qualification to Final Product quality assurance). 

Francisca F Gouveia, MSc in Biology, MSc in Pharmaceutical Engineering (TU Lisbon, 2010). She is a Senior MSAT Specialist and Partner at 4Tune Engineering Ltd, and responsible for the company’s ISO 9001:2008 Quality Management System. Ms Gouveia specialises in advanced PAT implementations and development of scale-independent strategies for biopharma applications.